lib.pci: Support short/canonical ("01:00.0") PCI addresses #669
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Adds these two functions:
pci.canonical(address) => returns address in short canonical format
pci.qualified(address) => returns address in fully qualified format
with the intention that users can specify either format and Snabb
Switch will prefer the canonical format internally but expand to the
qualified format when needed (e.g. for resolving sysfs paths).
This is a user visible change. The PCI selftest now shows canonical addresses:
$ sudo ./snabb snsh -t lib.hardware.pci
selftest: pci
pciaddress vendor device interface status driver usable
01:00.0 0x8086 0x10fb p1p1 down apps.intel.intel_app yes
01:00.1 0x8086 0x10fb p1p2 down apps.intel.intel_app yes
02:00.0 0x8086 0x10fb p2p1 down apps.intel.intel_app yes
02:00.1 0x8086 0x10fb p2p2 down apps.intel.intel_app yes
03:00.0 0x8086 0x10fb p3p1 down apps.intel.intel_app yes
03:00.1 0x8086 0x10fb - - apps.intel.intel_app yes
81:00.0 0x8086 0x10fb p5p1 down apps.intel.intel_app yes
81:00.1 0x8086 0x10fb p5p2 down apps.intel.intel_app yes
82:00.0 0x8086 0x10fb p6p1 down apps.intel.intel_app yes
82:00.1 0x8086 0x10fb p6p2 down apps.intel.intel_app yes
and software now accepts either format in parameters:
$ sudo SNABB_PCI0=01:00.0 SNABB_PCI1=0000:01:00.1 ./snabb snsh -t apps.intel.intel_app
selftest: intel_app
100 VF initializations:
Running iterated VMDq test...
test # 1: VMDq VLAN=101; 100ms burst. packet sent: 628,830
test # 2: VMDq VLAN=102; 100ms burst. packet sent: 1,745,985
test # 3: VMDq VLAN=103; 100ms burst. packet sent: 2,891,700
This change can potentially break software that depends on always
having qualified paths: that would now be considered a bug.
Explicitly expand device PCI addresses to their fully qualified representation before matching them with the patterns on the command line. This is to preserve the behavior of the packetblaster command line (e.g. accepting 0000: in address patterns) after the pci module adopted abbreviated canonical addresses in 27c1e6d and will return abbreviated names by default.
|
Seems reasonable. |
lukego
added a commit
to lukego/snabb
that referenced
this pull request
Dec 3, 2015
This was referenced Dec 8, 2015
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Proposed alternative to #657 after tomatoes were thrown in response to my more complicated proposal.
Adds these two functions:
pci.canonical(address)=> returns address in short canonical format (01:00.0)pci.qualified(address)=> returns address in fully qualified format (0000:01:00.0)with the intention that users can specify either format and Snabb Switch will prefer the canonical format internally but expand to the qualified format when needed (e.g. for resolving sysfs paths).
This is a user visible change. The PCI device scan and selftest now uses canonical addresses:
and software now accepts either format in parameters:
This change can potentially break software that depends on always having qualified paths: that would now be considered a bug. (This PR includes a fix for such a problem in packetblaster.)